Simulating financial time series using attention
This work addresses the challenge of generating realistic financial data for improved strategy evaluation, but it is incremental as it builds on existing GAN methods with attention enhancements.
The authors tackled the problem of simulating financial time series to augment limited real data for training trading strategies, by introducing two generative adversarial networks (GANs) with attention mechanisms. They tested these models on S&P 500 index and option data, showing that the attention-based GANs reproduced stylized facts and smoothed the autocorrelation of returns.
Financial time series simulation is a central topic since it extends the limited real data for training and evaluation of trading strategies. It is also challenging because of the complex statistical properties of the real financial data. We introduce two generative adversarial networks (GANs), which utilize the convolutional networks with attention and the transformers, for financial time series simulation. The GANs learn the statistical properties in a data-driven manner and the attention mechanism helps to replicate the long-range dependencies. The proposed GANs are tested on the S&P 500 index and option data, examined by scores based on the stylized facts and are compared with the pure convolutional GAN, i.e. QuantGAN. The attention-based GANs not only reproduce the stylized facts, but also smooth the autocorrelation of returns.